Error converting to GGUF

#1
by bartowski - opened

Getting this error:

ValueError: Can not map tensor 'layers.41.experts.66.w1.weight'

Using the --mistral-format arg with convert_hf_to_gguf.py, all other sizes converted without issue

Not sure you'd have the answer, but putting it here for discussion

gguf conversion might work with the git branch of transformers installed. I'd try it for you but it would take forever to download this model.

There is also an rc currently:

pip install transformers==5.0.0.rc03

ngxson has a draft PR up:

https://github.com/ggml-org/llama.cpp/pull/17730

Using it now, will know results in a couple hours

Sign up or log in to comment